tax avoidance and investment: distinguishing the …
TRANSCRIPT
TAX AVOIDANCE AND INVESTMENT: DISTINGUISHING THE EFFECTS OF
CAPITAL RATIONING AND OVERINVESTMENT
A Dissertation
by
MICHAEL MAYBERRY
Submitted to the Office of Graduate Studies of
Texas A&M University
in partial fulfillment of the requirements for the degree of
DOCTOR OF PHILOSOPHY
Approved by:
Chair of Committee, Connie Weaver
Committee Members, Thomas Omer
Anwer Ahmed
Paige Fields
Head of Department, James Benjamin
December 2012
Major Subject: Accounting
Copyright 2012 Michael Mayberry
ii
ABSTRACT
I examine the relation between tax avoidance and firm investment by drawing on
two capital market imperfections, adverse selection and moral hazard, to provide a link
between tax avoidance and investment. Firms experiencing capital rationing because of
adverse selection rely on internal resources to fund investment opportunities because of
costly external financing. Tax avoidance can provide additional cash-flows that may
alleviate capital rationing. Alternatively, tax avoidance can exacerbate problems of
moral hazard by facilitating managerial rent extraction in the form of overinvestment. I
find a positive relation between tax avoidance and investment suggesting effects of
either capital rationing or overinvestment. To distinguish between these two effects, I
examine how the relation between tax avoidance and investment varies in settings where
capital rationing or overinvestment is more likely to occur. My findings suggest that
firms rely on the cash savings from tax avoidance to alleviate capital rationing.
iii
DEDICATION
To God’s Grace, Calvin’s Five Points, and Double Imputation
iv
ACKNOWLEDGEMENTS
I would like to thank my Chair, Connie Weaver, and my dissertation committee,
Tom Omer, Anwer Ahmed, and Paige Fields. I would also like to thank workshop
participants at Texas Tech University and the University of Florida. I have also benefited
from helpful suggestions from Sean McGuire, Shiva Sivaramakrishnan, Brad Lawson,
and Bret Scott. On a personal note, the opportunities I have are a direct result of God’s
grace working through a number of people who I must acknowledge. I would not be in
the position to obtain a Ph.D. if it were not for my family. My mother taught me how
read and do arithmetic at an early age. She taught me study habits which I use until this
day. My father tutored me in mathematics and physics, usually after already having
worked a twelve hour day. Without stressed-out phone calls to my sister, often at
unreasonable hours, I doubt I could have maintained sanity. God has richly blessed with
a number of friends. Christopher Bounds, Scott Hansen, Casey Carmichael, Adam
Friudenberg, Dave Jones, and Christopher Duke all stood beside me during dark days,
prayed for me, and preached the truth to me. My officemate, Brad Lawson, never let me
get discouraged and give up. Finally, I owe a debt of gratitude to MikeBrown, Tim
Tebow, and Ronald Reagan for unparalleled brilliance, two national championships, and
winning the Cold War and the Tax Reform Act of 1986, respectively.
v
TABLE OF CONTENTS
Page
ABSTRACT ................................................................................................................. ii
DEDICATION ............................................................................................................. iii
ACKNOWLEDGEMENTS ......................................................................................... iv
TABLE OF CONTENTS ............................................................................................. v
CHAPTER
I INTRODUCTION ................................................................................... 1
II HYPOTHESIS DEVELOPMENT .......................................................... 9
Investment with Imperfect Capital Markets ............................. 9
Tax Avoidance ......................................................................... 11
Distinguishing Capital Rationing and Overinvestment ............ 14
III VARIABLE DEFINITION AND METHODOLOGY ........................... 19
Model Development ................................................................. 19
Controls .................................................................................... 26
IV SAMPLE SELECTION .......................................................................... 30
V RESULTS ................................................................................................ 32
Univariate Statistics .................................................................. 32
Correlations .............................................................................. 32
Multivariate Results ................................................................. 33
Determining the Source of the Positive Relation ..................... 35
VI ROBUSTNESS TESTS ........................................................................... 38
VII CONLUSION .......................................................................................... 43
REFERENCES ............................................................................................................. 45
vi
Page
APPENDIX A ............................................................................................................. 51
APPENDIX B ............................................................................................................. 54
1
CHAPTER I
INTRODUCTION
In perfect capital markets, managers should be indifferent between internal and
external resources to finance their investment (Modigliani and Miller 1958).
Nevertheless, numerous studies find that internal resources are significant predictors of
firms’ investment (Almeida and Campello 2007; Fazzari et al. 1988; Lamont 1997).
Theorists posit that two capital market imperfections can lead to firms relying on internal
resources to fund investment: adverse selection and moral hazard (Stein 2003). Adverse
selection occurs because managers possess greater information about the true value of
their firm and opportunistically obtain outside financing when it benefits current
shareholders (Myers 1977; Myers and Majluf 1984). Potential new shareholders are
aware of the information asymmetry problem and ration the capital they offer to
managers, thus making external financing more costly. An increase in the cost of
external financing encourages firms to use internal funds to finance investments.
Moral hazard refers to managers’ willingness to extract rents from the firm due to
the difference in utility functions between managers and shareholders (Jensen and
Meckling 1976) and may lead managers to rely on internal resources because they are
reluctant to expose their rent extraction to outside financiers by accessing external
capital markets (Harris and Raviv 1990; Jensen 1986; Stein 2003; Stulz 1990). Moral
hazard can lead managers to overinvest, where managers make unprofitable investments
to increase the size of the firm for managers’ personal gain (Jensen 1986).
2
Because of this reliance on internal resources, managers may search for means of
financing investments in addition to cash flow generated by its core operations. For
example, a reduction in firms’ explicit tax liability (henceforth tax avoidance) increases
a firm’s after-tax cash flow and therefore increases the internal resources available for
investment. Tax avoidance therefore can influence the level of firms’ investment in
circumstances in which firms must rely on internal resources. In this paper, I examine
whether tax avoidance and investment are related and investigate whether adverse
selection-induced capital rationing (henceforth capital rationing) or moral hazard
influence that relation.
Both capital rationing and overinvestment predict a positive relation between tax
avoidance and investment. In the case of capital rationing, managers use the cash savings
from tax avoidance to help mitigate a lack of internal resources to fund investment. All
else equal, firms with more internal resources can rely less on costly external financing
and forgo fewer profitable investment opportunities (Fazzari et al. 1988; Myers and
Majluf 1984). However, in addition to increasing the internal resources of the firm, tax
avoidance also increases the probability of moral hazard by increasing information
asymmetry between managers and investors (Balakrishnan et al. 2011; Desai and
Dharmapala 2009; Hope et al. 2012). This information asymmetry decreases the ability
of investors to detect unwarranted rent extraction (Jensen and Meckling 1976).1 Because
tax avoidance increases information asymmetry, Desai and Dharmapala (2009) argue
1 Managers are unable to maintain a higher level information asymmetry with taxing authorities than they
have with their shareholders. The IRS is able to observe public disclosures made by corporate managers to
shareholders and apparently use such disclosures during the audit process (Mills 1998).
3
that tax avoidance lowers the marginal cost of rent extraction. In the case of
overinvestment, managers exploit the increase in information asymmetry from tax
avoidance and use the cash savings of tax avoidance to invest beyond what is optimal for
shareholders for reputational or job security reasons, resulting in greater investment.
I find that tax avoidance is positively associated with both firm-level investment
and unexpected investment, which controls for firms’ investment opportunities. On
average, firms’ invest between 10.3 and 17.9-cents for every dollar of cash flows saved
through tax avoidance.2 This finding is consistent with the relations predicted by both
capital rationing and overinvestment.
To distinguish whether the positive relation between tax avoidance and
investment is evidence of the benefits of alleviating capital rationing or the exacerbation
of overinvestment, I follow recent finance research by investigating how the positive
relation between internal resources and investment varies cross-sectionally (Babenko et
al. 2011; Hadlock 1998; Ozbas and Scharfstein 2010; Richardson 2006). More
specifically, I distinguish between capital rationing and overinvestment by investigating
the relation between tax avoidance and investment across different levels of managerial
ownership, relative internal resources, and volatility in tax avoidance.
Adverse selection, and therefore capital rationing, increases with firms’
managerial ownership because managers internalize more of the benefits of
2 While the amount firms invest from tax avoidance might appear low, the investments associated with tax
avoidance are approximately equal to the investments associated with pre-tax cash flows, suggesting that
firms do not behave radically different when investing the cash savings from tax avoidance. My research
design does not allow me to determine how firms spend remaining 89.7-cents to 82.1-cents of cash savings
from tax avoidance. However, a portion of the remaining cash savings from tax avoidance likely recoups
the up-front costs associated with tax avoidance. My estimates of the cash savings from tax avoidance do
not incorporate the legal fees, accountants fees, and organizational costs associated with tax avoidance.
4
opportunistically seeking external financing (Hadlock 1998; Ozbas and Scharfstein
2009). However, overinvestment decreases with managerial ownership because
managers internalize more of the costs of overinvestment (Hadlock 1998; Jensen 1986).
Thus, a finding of a significantly more positive relation between tax avoidance and
investment among firms with a high level of CEO ownership is consistent with firms
that experience a higher level of capital rationing alleviating its effect on investment
through tax avoidance. A finding of a significantly less positive association between tax
avoidance and investment among firms with a high level of CEO ownership suggests
managers refrain from overinvestment when they bear a greater portion of its costs and
would suggest that overinvestment is more likely to be the source of the positive
relation.
Firms with low investment opportunities and positive free cash flow have greater
internal resources relative to investment opportunities and are less likely to experience
financing deficits than firms with more investment opportunities and negative free cash
flow (Hadlock 1998; Hoshi et al. 1991). 3
To the extent that capital rationing affects
firms’ investment, firms with greater relative internal resources would not need to rely as
heavily on tax avoidance to fund all investment opportunities as firms with lower
relative internal resources. Therefore, if firms engage in tax avoidance to increase their
internal resources and alleviate capital rationing, the relation between tax avoidance and
investment will be less positive among firms with greater relative internal resources and
more positive among firms with high investment opportunities and negative free cash
3 Free cash flow is cash flow beyond what is necessary to finance positive net present value (NPV)
investments (Jensen 1986; Richardson 2006)
5
flow (less relative internal resources). Overinvestment, on the other hand, should be
more salient when relative internal resources are greater because managers’ preference
for larger firms is independent of investment opportunities (Hoshi et al. 1991) and
positive free cash flow provides an opportunity for managers to grow the size of the firm
(Jensen 1986). A significantly more positive relation between tax avoidance and
investment among firms with greater relative internal resources suggests overinvestment
as the likely source of the positive relation.
Firms with higher cash flow volatility are more likely to experience financing
deficits and forgo investment because of capital rationing (Minton and Schrand 1999).
Firms with volatile tax avoidance strategies are similarly likely to suffer shortfalls in
cash flows related to tax avoidance and are less able to rely on tax avoidance to
consistently fund investments. To the extent that firms rely on the cash flow from tax
avoidance to overcome capital rationing and fund investments, the relation between tax
avoidance and investment will be less positive for firms with volatile tax avoidance
strategies and more positive for firms with less volatile tax avoidance strategies. Prior
evidence suggests managers are likely to overinvest unexpected funds from cash
windfalls, indicating overinvestment is more likely to occur in settings of volatile tax
avoidance (Blanchard et al. 1994). A significantly more positive relation between tax
avoidance and investment among firms with volatile tax avoidance strategies suggests
overinvestment as the likely source of the positive relation.
I find that capital rationing is more likely to be the source of the positive relation
between tax avoidance and investment, rather than overinvestment. Firms with higher
6
levels of managerial ownership have a significantly more positive association between
tax avoidance and investment than firms with lower levels of managerial ownership.
Also, firms with greater relative internal resources have a less positive association
between tax avoidance and investment than firms with less relative internal resources.
Finally, firms with volatile tax avoidance have a less positive association between tax
avoidance and investment compared to firms with less volatile tax avoidance strategies.
Overall, my findings are consistent with the view that, on average, tax avoidance is a
source of internal financing for firms.
My findings are important for several reasons. First, understanding the
consequences of tax avoidance is important to shareholder wealth, especially if forgone
investment opportunities or overinvestment are costly to the firm (Minton et al. 2002;
Minton and Schrand 1999). Second, the consequences of tax avoidance are important to
policymakers as they try to understand the costs and benefits associated with curtailing
tax avoidance. My findings suggest that curtailing tax avoidance may decrease
investment. Third, my results are important to managers seeking to overcome financing
constraints and optimize investment and suggest that tax avoidance can, on average, be
used as a form of internal financing.
I contribute to several areas of accounting research. Previous literature examines
the measurement (Dyreng et al. 2008; Frank et al. 2009) and determinants (Badertscher
et al. 2010; Chen et al. 2010; Gaertner 2010; Higgins et al. 2011; McGuire et al. 2011b;
Minnick and Noga 2010; Rego and Wilson 2012; Wilson 2009) of tax avoidance. It is
only recently that researchers have examined the financial effects of tax avoidance,
7
including leverage (Graham and Tucker 2006), cash holdings (Dhaliwal et al. 2011),
firm value (Desai and Dharmapala 2009; Koester 2011; Wang 2010; Wilson 2009), and
investor reaction (Frischmann et al. 2008; Hanlon and Slemrod 2009; Kim et al. 2010). I
contribute to this line of literature by examining the effect of tax avoidance on
investment, providing real effects to support the prior financial policy and price effects.
Blaylock (2011) and Khurana et al. (2011) find conflicting evidence on whether tax
avoidance is related to firm’s investment.4 I am able to provide additional evidence on a
positive relation between tax avoidance and investment that may help to reconcile the
results of these studies. Second, I contribute to the corporate governance view of tax
avoidance, Desai and Dharmapala (2006) theorizes a moral hazard framework for a
manager’s tax avoidance decisions. I expand this by examining tax avoidance from the
perspective of adverse selection, another prevalent capital market imperfection that can
affect investment through capital rationing. Third, I contribute to the corporate finance
literature by providing additional evidence on the capital market imperfection which
links investment to internal resources (Stein 2003; Hadlock 1998). Lastly, I contribute to
the literature on firms’ responses to financing constraints that considers financial
reporting quality (Biddle et al. 2009; Biddle and Hilary 2006), blockholder ownership
(Allen and Phillips 2000), and cross-subsidies within the firm (Hadlock et al. 2001) as
means of alleviating financing constraints. I propose another firm-level response for the
alleviation of financing constraints, tax avoidance.
4 Both Blaylock (2011) and Khurana et al. (2011) consider tax avoidance from a moral hazard perspective
arguing that a positive relation between tax avoidance and investment results from moral hazard problems.
In addition to evaluating the moral hazard argument, I also consider adverse selection as an explanation
for any positive relation.
8
The remainder of the paper is structured as follows. In the next section, I discuss
prior literature on investment and tax avoidance and develop my hypotheses. In Section
3, I define variables and describe the methodology. Section 4 details my sample while
Section 5 discusses the univariate and multivariate. In Section 6, I perform robustness
tests. Section 7 concludes.
9
CHAPTER II
HYPOTHESIS DEVELOPMENT
Investment with Imperfect Capital Markets
When capital markets contain perfect information, internal and external resources
are perfect substitutes for financing investments (Modigliani and Miller 1958). Under
this view of the firm, managers invest resources until the marginal benefit of investment
equals the marginal cost and investment is driven by investment opportunities (Hayashi
1982; Tobin 1969). If firms experience a financing deficit, where investment
opportunities exceed internal resources, external resources can be obtained without
excessive cost. However, prior research documents a strong relation between firms’
internal resources and the level of investment indicating that managers must, at least
partially, rely on internally generated cash flow for financing (Lamont 1997; Richardson
2006; Stein 2003). Fazzari et al. (1988) demonstrate that firm-level cash flow is a
significant determinant of investment, even after controlling for investment
opportunities. Alternatively, Blanchard et al. (1994) find that firms increase their
investment in response to cash windfalls from favorable legal settlements, despite low
investment opportunities.
Two different capital market imperfections can lead to this positive relation
between internal resources and investment: adverse selection and moral hazard (Stein
2003). Adverse selection arises because of information asymmetry between the manager
and outside financiers (Myers and Majluf 1984; Stiglitz and Weiss 1981). Myers and
10
Majluf (1984) model adverse selection for equity financing and suggest that managers
exploit their informational advantage by issuing equity when they have information that
the firm is overvalued.5 As a result, the capital obtained from new shareholders benefits
current shareholders at the expense of new shareholders. In equilibrium, potential
shareholders are aware of this transfer of wealth and increase their required return by
reducing the quantity of equity capital available to managers, a phenomenon described
as capital rationing. In Myers and Majluf (1974), management is unable to obtain outside
financing for all positive net present value (NPV) projects because of capital rationing.
As a result, firms rely on internal resources to fund investments and may forgo positive
NPV investments when they face a financing deficit.6
In contrast to capital rationing, which arises because of the increased cost of
external financing, moral hazard arises because managers take actions to benefit
themselves at the expense of all shareholders. Jensen and Meckling (1976) model moral
hazard as the difference in utility functions between managers and shareholders.
Managers’ interests are not perfectly aligned with shareholders’ interests because
managers do not bear the full cost of suboptimal behavior. Managers’ preference for
firm size or complexity may be different than shareholders because their private benefits
of control increase with firm size (Hart and Moore 1995; Jensen 1986). As a result,
5 In terms of debt financing, Myers (1977) notes that managers are able to either distribute debt proceeds
to shareholders or invest in riskier assets to benefit shareholders at the expense of debtholders.
Anticipating this misappropriation, debtholders increase their required return by decreasing the amount of
debt capital offered to managers. 6 Adverse selection and capital rationing are related concepts in that adverse selection induces capital
rationing (Hadlock 1998; Hubbard 1998; Rauh 2006; Stein 2003; Vogt 1994). Adverse selection arises
because of managers exploiting new shareholders and debtholders. Capital rationing arises because of
shareholders’ and debtholders’ response to adverse selection, namely, the increased cost of external
financing. All else equal, firms with larger adverse selection problems face greater capital rationing.
11
managers invest in negative NPV projects to increase the size of the firm, a phenomenon
referred to as overinvestment. In the case of overinvestment, managers do not seek
external financing because it would discipline their self-interested behavior (Harris and
Raviv 1990; Jensen 1986; Stein 2003; Stulz 1990). Therefore, firms rely on internal
resources to fund investments.
Tax Avoidance
Tax avoidance is “the reduction of explicit taxes” (Hanlon and Heitzman 2010).
Following Hanlon and Heitzman (2010), I consider tax avoidance to be a continuum that
ranges from objectively legal strategies (e.g., investments in municipal bonds) to those
of uncertain legality (e.g., tax shelters). 7 Traditionally, researchers have examined tax
avoidance in the context of a trade-off where the benefits of reducing taxes are balanced
against financial reporting and regulatory costs (Shackelford and Shevlin 2001; Scholes
et al. 2008). Under this traditional view, a firm’s level of tax avoidance balances
marginal benefit and marginal cost and is therefore beneficial to shareholders. However,
Desai and Dharmapala (2006) discuss tax avoidance in the context of a moral hazard
framework, in which tax avoidance is not always optimal for shareholders because it
may facilitate management’s self-interested behavior. The traditional and the moral
hazard views of tax avoidance provide theoretical links to a firm’s investment.
7 While it is difficult to operationally distinguish tax avoidance and tax evasion, tax evasion is the
fraudulent and illegal reduction of explicit taxes whereas extreme forms of tax avoidance involve tax
benefits that are uncertain and might exploit tax regulations in a manner unintended by taxing authorities
(Frank et al. 2009; Hanlon and Heitzman 2010).
12
By definition, tax avoidance increases firms’ after-tax cash flows.8 All else equal,
firms with higher levels of tax avoidance pay less (as a percentage of income) to the
government than firms with lower levels of tax avoidance. Firms can use the increased
after-tax cash flow from tax avoidance to alleviate the effects of capital rationing and
maintain a higher level of investment compared to firms with a lower level of tax
avoidance.9 A dollar kept from taxing authorities is therefore an additional dollar of
internal resources and, therefore, an additional dollar available for investment. The
income shifting literature provides limited evidence of this positive relation between tax
avoidance and investment. Oversech (2009) provides evidence that foreign firms invest
more in Germany as their ability to shift profits outside of Germany increases. Grubert
(2003) similarly finds an increase in R&D investment among multinationals that shift
income to low tax jurisdictions. Both Oversech (2009) and Grubert (2003) attribute their
findings to tax avoidance increasing firm profitability, a form of internal resources.
However, the extent to which Oversech’s (2009) and Grubert’s (2003) findings apply to
US firms as well as to tax avoidance that is not income shifting is an open, empirical
question. Evidence from the valuation literature is also consistent with capital rationing.
Koester (2011) and Wang (2010) find that tax avoidance is, on average, positively
valued by the market. A positive valuation could imply that tax avoidance alleviates
capital rationing and allows firms to invest in positive NPV projects. If firms use the
8 This assumes that implicit taxes do not equal tax preferences (Jennings et al. Forthcoming). Implicit
taxes are decreases in pre-tax returns of tax-favored assets. Jennings et al. (Forthcoming) find that after
1986, implicit taxes do not equal tax preferences. To the extent that tax avoidance leads to implicit taxes,
my tests are biased against finding results. 9 Tax avoidance can be used to reduce the effects of capital rationing on investment but does not directly
alleviate adverse selection problems. It is unclear whether future shareholders benefit from current period
tax avoidance.
13
cash flow savings from tax avoidance to mitigate the effects of capital rationing, there
should be a positive relation between tax avoidance and investment.
Desai and Dharmapala (2006) consider tax avoidance from an agency cost
perspective, where managers are able to use tax avoidance to facilitate their rent
extraction. Because the value of tax avoidance is decreasing in the probability of
detection (Allingham and Sandmo 1972), there must be a level of obfuscation for tax
avoidance to be beneficial to shareholders.10
Balakrishnan et al. (2011) and Hope et al.
(2012) provide empirical confirm of Desai and Dharmapala’s (2009) theory and find that
tax avoidance is positively associated with information asymmetry among shareholders.
The greater the obfuscation, the lower the marginal cost of manager’s rent extraction
because monitoring is one primary constraint on a manager’s rent extraction (Desai and
Dharmapala 2006; Jensen 1986; Jensen and Meckling 1976). If tax avoidance increases
the information asymmetry associated with public disclosures for both shareholders and
taxing authorities then it likely not only reduces detection risk but also hides rent
extraction from shareholders. Empirical evidence supports the moral hazard view of tax
avoidance (Desai and Dharmapala 2006; Koester 2011; Wang 2010; Wilson 2009). Tax
avoidance may, therefore, increase a manager’s ability to conceal from shareholders the
consumption of internal resources for personal gain. Managers can use tax avoidance to
facilitate overinvestment because tax avoidance provides both increased after-tax cash
flows and the ability to conceal self-interested actions from shareholders. To the extent
that mangers overinvest the cash savings, there should be a positive relation between tax
10
This obfuscation is not only related to taxing authorities, but also to shareholders. Taxing authorities are
aware of financial statement disclosures and employ them in the audit process (Mills 1998).
14
avoidance and investment. Given the similarly positive predictions from capital rationing
and moral hazard and the inconsistency of prior empirical evidence, I hypothesize in the
alternative form:
H1: Tax avoidance is significantly and positively associated with the level of
firm’s investment.
A positive relation would suggest managers use the cash savings from tax
avoidance to alleviate capital rationing or that managers exploit the information
asymmetry associated with tax avoidance by using the cash savings from tax avoidance
to overinvest. I discuss how to distinguish capital rationing and overinvestment in the
following section.
Distinguishing Capital Rationing and Overinvestment
A simple association test between tax avoidance and investment cannot
distinguish adverse selection’s capital rationing and moral hazard’s overinvestment
because both predict a positive association. Previous research establishing a relation
between tax avoidance and investment has exclusively relied on overinvestment as
motivation (Khurana et al. 2011; Blaylock 2011). However, as Stein (2003), Bergstresser
(2006), and Hadlock (1998) argue, capital rationing can also produce a positive relation
between internal resources and investment. To determine if a positive relation between
tax avoidance and investment is the result of capital rationing or overinvestment, I
further examine the relation in three different settings.
First, adverse selection and moral hazard make different assumptions about the
behavior of managers. Realizing this difference between capital rationing and
15
overinvestment, Hadlock (1998) and Ozbas and Scharfstein (2010) partition their
samples on managerial ownership to determine if a positive relation between internal
resources and investment is due to capital rationing or overinvestment. Managerial
ownership exacerbates adverse selection because CEOs internalize a greater portion of
the benefits from issuing overvalued securities. Therefore, managerial ownership
increases the adverse selection problem, thus increasing capital rationing, and increasing
firm’s reliance on the cash savings from tax avoidance. To the extent that capital
rationing is the source of a positive relation between tax avoidance and investment, then
the association will be significantly greater when managerial ownership is high.
Managerial ownership, on the other hand, alleviates moral hazard because CEOs
internalize a greater portion of the costs associated with overinvestment. To the extent
that overinvestment is the source of the positive association between tax avoidance and
investment, then the association should be significantly less positive when managerial
ownership is high. Given the different predictions offered by capital rationing and
overinvestment, I offer a hypothesis in the null form:
H2: The association between tax avoidance and the level of firm’s investment is
not significantly different between firms with high CEO ownership and firms
without high CEO ownership.
Second, the level of internal resources relative to investment opportunities affects
capital rationing and overinvestment differently. Hoshi et al. (1991) and Richardson
(2006) partition on Tobin’s Q and free cash flow, respectively, to take advantage of the
different effects internal resources have on capital rationing and overinvestment in order
16
to identify the source of a positive relation between internal resources and investment.
All else equal, capital-rationed firms with more investment opportunities and insufficient
internal resources are more likely to experience a financing deficit and would be more
likely to forgo investments because of the excessive cost of external financing. Firms
with less relative internal resources would therefore benefit the most from tax
avoidance.11
If tax avoidance and investment are related due to capital rationing, then the
association between tax avoidance and investment should be significantly less positive
among firms with greater relative internal resources as these firms are most likely to
forgo investments. On the other hand, overinvestment is most likely to occur in firms
with greater relative internal resources (poor investment opportunities and positive free
cash flow). Managers’ incentive to overinvest does not decrease with investment
opportunities (Hoshi et al. 1991; Jensen 1986). Moreover, managers with positive free
cash flow have the opportunity to invest beyond the optimal level (Jensen 1986;
Richardson 2006). All else equal, firms with greater relative internal resources are more
likely to invest suboptimally. If tax avoidance and investment are related due to
overinvestment, then the association should be greater among firms with high relative
internal resources. I hypothesize:
H3: The association between tax avoidance and the level of firm’s investment is
not significantly different between firms with greater relative internal resources
and firms with less relative internal resources.
11
Greater relative internal resources do not impact adverse selection, directly, but exacerbate the effect
capital rationing has on investment by increasing the need for costly external financing.
17
Finally, the volatility of tax avoidance has different effects on capital rationing
and overinvestment. Firms experiencing capital rationing must seek internal forms of
financing. When forms of internal financing are volatile, firms are more likely to
experience financing deficits and, because of capital rationing, forgo investment (Minton
and Schrand 1999). 12
If capital rationing contributes to the positive relation between tax
avoidance and investment, volatile tax avoidance strategies will be less able to finance
investment because of the risk that cash flows from tax avoidance do not materialize.13
To the extent that tax avoidance and investment are related due to capital rationing, then
the association between tax avoidance and investment should be significantly less
positive among firms with volatile tax avoidance strategies compared to firms with less
volatile tax avoidance strategies. While volatile tax avoidance is less likely to finance
investments under capital rationing, volatile cash flows from tax avoidance are more
likely to fund overinvestment. Blanchard et al. (1994) argue that firms with large cash
windfalls are likely to overinvest as the unexpected cash flow provides managers an
opportunity to grow the firm. Controlling for the level of tax avoidance, more volatile
tax avoidance strategies are likely to reflect one-time cash windfalls from short-term tax
strategies or settlements with taxing authorities. If overinvestment influences the positive
relation between tax avoidance and investment, then the association will be significantly
12
Interacting tax avoidance with the volatility of operating cash flows would not aide in distinguishing
capital rationing and overinvestment. As discussed above, volatile operating cash flows lead firms that are
experiencing capital rationing to forgo investment. However, to the extent that volatile operating cash
flows represent one-time cash flows, then overinvestment is also likely to be exacerbated (Bates 2005;
Blanchard et al. 1994). 13
The volatility of operating cash flows, not the volatility of tax planning, is likely to be the primary
reason a firm forgoes investment due to capital rationing. However, firms that are already experiencing
capital rationing are the most likely to seek stable (i.e. less volatile) forms of internal financing to alleviate
the effects of capital rationing. Thus, I interact tax avoidance with the volatility of tax avoidance.
18
more positive among firms with more volatile tax avoidance. Given the conflicting
expectations, I hypothesize:
H4: The association between tax avoidance and the level of firms’ investment
does not significantly vary with the volatility of firms’ tax avoidance.
19
CHAPTER III
VARIABLE DEFINITION AND METHODOLOGY
I test my hypotheses using two different methods. First, I test the association
between tax avoidance and investment, controlling for other documented determinants of
investment. Second, I distinguish whether capital rationing and overinvestment are the
likely source of a positive association by investigating how the association varies
between firms with high managerial ownership, greater relative internal resources, and
the volatility of tax avoidance.
Model Development
I use Model (1) as the primary test for my first hypothesis. Model (1) is an OLS
regression of firms’ investment, defined as either as investment or unexpected
investment, on my measure of tax avoidance and other control variables. I estimate
Model (1) over the entire sample, adjust standard errors for firm and year clustering, and
include industry fixed effects (bj).
Ii,t = b0 + b1 negCETR5i,t-1 + Controls’k,t-1 + bj + ei,t (1)
where:
negCETR5i,t-1 = My proxy for tax avoidance, defined in Equation (3)
Controls’k,t-1 = Control variables detailed below
bj = Industry fixed-effects based on Fama-French 48 definitions
The dependent variable of Model (1) is either Invest or UnExp_Invest.
Investment (Invest) is the sum of capital expenditure (CAPX), research and development
20
(XRD), and acquisitions (AQC) less the sale of property, plant, and equipment (SPPE)
and scaled by lagged total assets (AT) in accordance with Biddle et al. (2009).14
Invest
has the advantage of not making assumptions about firms’ investment functions but
suffers the disadvantage of not controlling for growth opportunities. Therefore, my
second measure of investment, UnExp_Invest, is the residual from an industry-year
regression of investment (Invest) on investment opportunities (Opportunities) (Model
(2)). UnExp_Invest controls for how a firm’s industry invests based on its investment
opportunities.15
Model (2) follows directly from Verdi (2006) as well as conceptually
with the two-stage approach used by Richardson (2006). This measure allows me to
distinguish between two firms’ investment levels with different investment opportunities
that are in the same industry.
Investi,t = a0 + a1 Opportunitiesi,t-1 + ei,t (2)
Model (2) is based on the idea that when capital markets are perfect and
frictionless, a firm’s investment should be a function of investment opportunities
(Hayashi 1982; Hubbard 1998; Tobin 1969). Traditionally, a firm’s optimal investment
is a linear function of its Tobin’s Q.16
However, recent research acknowledges that Q’s
denominator, the book value of assets, can be biased by the accounting system’s
conservatism and suggests sales growth as an alternate proxy for a firm’s investment
14
Following Biddle et al. (2009), I also analyze only capital expenditures (CAPX) and research and
development (XRD) as measures of firm-level investment. Results are inferentially similar for both capital
expenditures and research and development forms of investment. 15
I continue to use industry fixed-effects to control for the industry-average value of tax avoidance and
control variables for firms in my sample. 16
Tobin’s Q is the ratio of a firm’s market value of assets to book value of assets ((AT +
(PRCC_F*CSHO) – CEQ)/ AT).
21
opportunities (Biddle et al. 2009; Richardson 2006).17
I present results using sales
growth as a proxy for investment opportunity to avoid potential biases associated with
using Tobin’s Q, consistent with Biddle et al. (2009) and Verdi (2006). However, I
acknowledge that sales growth is also not a perfect proxy for investment opportunities.
Therefore, I remove observations where residuals have conflicting signs across the
Tobin’s Q and sales growth specification (6,254 firm-years).18, 19
The variable of interest in Model (1) is negCETR5, which is the five-year cash
effective tax rate from Dyreng et al. (2008) multiplied by negative one.20
Using cash
taxes paid has the advantage of capturing both permanent and temporary forms of tax
avoidance without the effects of GAAP tax accruals, such as tax cushions or valuation
17
The potential bias in Tobin’s Q can produce spurious results in my study for two reasons. First, financial
accounting introduces conservatism to the measure of Tobin’s Q and thus could potentially create spurious
correlations. I want to avoid spurious correlations with tax avoidance measures because there is evidence
that tax avoidance and financial reporting quality are positively correlated (Frank et al. 2009). Second,
research and development expenditures can qualify for the R&D tax credit, which lowers a firm’s ETR
and, through the expensing of R&D, increases a firm’s Tobin’s Q. This also can introduce a spurious
correlation because the level of research and development expenses is likely mechanically related to
Tobin’s Q. 18
Comparing the relation between investment and sales growth within an industry allows for a benchmark
across firms but the OLS specification of Model (1) forces deviations from expected investment to average
to zero in every industry-year. Verdi (2006) notes that deviations from expected investment can be
pervasive across industries, suggesting that an OLS residual may be an inappropriate measure because
they mechanically average to zero. In an untabulated robustness test, I allow for a given industry to
experience pervasive deviations from expected investment by adding the difference between the industry-
year average investment and total sample average investment to the residual from Model (1). This
adjustment assumes that, in the long run, all firms invest optimally and allows for a given industry-year to
broadly experience incentives to deviate from expected investment. Results are qualitatively similar. 19
Results are qualitatively similar when I include firms with conflicting signs across Tobin’s Q and sales
growth. In addition, my results hold when I simply use the traditional Tobin’s Q variable as a proxy for
investment opportunities. 20
In untabulated analysis, I use firms’ five-year current ETR to measure tax avoidance. Using the current
ETR, I find a positive relation between tax avoidance and investment among firms that invest above their
expected investment level. I also find significant and positive interactions between current ETRs and CEO
ownership. Both findings are consistent with firms using tax avoidance to alleviate capital rationing and
not overinvestment. I present results using negCETR5 as this variable is arguably more highly correlated
with the cash savings from tax avoidance than current ETR. As firms fund investments with cash and not
earnings, negCETR5 appears a stronger measure for testing my hypotheses. In addition, financial
statement accruals, such as valuation allowances or tax cushions, do not influence cash ETRs but
positively bias current ETR measures.
22
allowances. However, cash taxes may suffer from a mismatching problem because
estimated tax payments for future periods and payments of back taxes are included in the
variable. Therefore, I accumulate cash taxes paid over the period t-5 to t-1 and scale by
pretax income net of special items over the same time period to minimize this
mismatching problem. To ease interpretation, I multiply negCETR5 by negative one so
that it is increasing in tax avoidance. It is important to note that the measurement of
negCETR5 precedes investment (i.e. the accumulation of numerator and denominator
ceases in year t-1) and does not overlap with the year that investment occurs (year t).
Current period investment and current period tax avoidance are mechanically related
because of accelerated tax depreciation and the research and development tax credit.
Measuring tax avoidance before the investment occurs minimizes the likelihood that
results are due to this mechanical association.21
negCETR5 = (-1)*[ TXPD / (PI-SPI)] (3)
While use of effective tax rates (ETRs) is common in accounting literature, one
notable drawback of ETRs is their inability to provide a dollar-value quantification of
the cash of the savings from tax avoidance. Therefore, I use Equation (4) in order to
provide a quantification of the cash savings from tax avoidance. Equation (4) assumes
that absent tax avoidance strategies, firms pay the US federal statutory tax rate of 35%.22
21
In a robustness test, I measure tax avoidance two periods prior to investment to further minimize the risk
of a mechanical association (i.e. I accumulate from year t-6 to t-2). Results are inferentially similar. 22
Even though the US Federal statutory tax rate is 35%, firms may be taxed at different rates if their
earnings are only taxable in foreign jurisdictions. The extent to which firms do not save at a rate of 35%
on tax avoidance activities introduce noise in my estimates of Tax Cash Savings and biases against my
finding results as my estimate of tax cash savings would deviate from the actual tax cash savings, reducing
the explanatory power of my estimate (Hanlon 2003).
23
I then multiply the difference between the statutory tax rate and a firm’s CETR5 by a
firm’s pretax income to transform the cash savings from a scaled value to a dollar value.
Tax Cash Savingsi,t-1 = [(.35-CETR5 i,t-1)*PI i,t-1]/ATi,t-1 (4)
I then include Tax Cash Savings in Equation (5) to estimate the portion of tax
cash savings that firms invest, on average. Equation (5) is a modified version of
Equation (1) and includes pretax cash flows (PTCF) as well as firm (bi) and year (bt)
fixed-effects.23
Pretax cash flows are operating cash flows (OANCF) plus cash taxes paid
(TXPD) scaled by lagged total assets. The firm and year fixed-effects allow me to
interpret the coefficient on Tax Cash Savings as the amount that firms invest from one
additional dollar of cash savings from tax avoidance controlling for the firm and year’s
average Tax Cash Savings and Invest.
Investi,t = b0 + b1 Tax Cash Savingsi,t-1 +b2PTCF i,t-1 +Controls’k,t-1 + bi +bt+ ei,t (5)
In my second set of tests, I provide evidence on whether capital rationing or
overinvestment is, on average, more likely to be the source of the positive association. I
use Model (1) and interact negCETR5 with three variables: managerial ownership,
relative internal resources, and the volatility of tax avoidance. All three variables provide
unique directional-predictions for capital rationing or overinvestment.
First, I obtain CEO ownership from Compustat’s Execucomp database. Then, I
sort firms into quintiles based on CEO ownership. I create an indicator variable,
HighCEO, which equals one if the firm is in the top quintile of CEO ownership and zero
23
I only use Invest as the dependent variable for Equation (5) to match the cash inflow of tax avoidance
with the cash outflow of investment. Because Invest represents the entire cash outflow of investment for a
given firm-year and UnExp_Invest only represents a portion of the investment outflow, estimating a
regression with UnExp_Invest as the dependent variable may underestimate the magnitude of investment
associated with the cash savings from tax avoidance.
24
otherwise. I interact HighCEO with my tax avoidance measure. A positive coefficient on
the interaction suggests that firms that avoid taxes and have high CEO ownership invest
more than firms that avoid taxes and have low CEO ownership, consistent with capital
rationing being the source of the positive relation. A negative coefficient on the
interaction suggest that as CEO’s incentives are increasingly aligned with shareholders,
investment levels decrease among firms that avoid taxes, consistent with overinvestment
being the source of the positive relation.
Second, I partition my sample on firms’ internal resources relative to investment
opportunities. I follow Hoshi et al. (1991) and Richardson (2006) by measuring firms’
investment opportunities using Tobin’s Q and the sign of free cash flow for internal
resources. I follow Richardson (2006) by defining free cash flow as the sum of operating
cash flows (OANCF), research and development (XRD), less depreciation (DPC) and the
predicted investment value from Model (2). I measure relative internal resources based
on the dimensions of free cash flow and investment opportunities. I create an indicator
variable (High_RIR) that equals one for firms with positive free cash flow and with
Tobin’s Q in the bottom two quintiles of its industry-year.24
Firms with high relative
internal resources have cash flow in excess of expect investment and low investment
opportunities for their industry and time period. I interact High_RIR with my tax
avoidance measure. A negative coefficient on the interaction suggests that firms that
avoid taxes and have high relative internal resources investment opportunities rely less
24
In untabulated analysis, I also interact tax avoidance with positive free cash flow and low values of
Tobin’s Q separately. Results for Tobin’s Q are inferentially similar to High_RIR. The interaction between
tax avoidance and positive free cash flow is insignificant.
25
on tax avoidance to fund investments than firms that avoid taxes and have low relative
internal resources, consistent with firms with a high level of preexisting internal
resources relative to investment opportunities relying on tax avoidance less to alleviate
capital rationing and fund investment. A positive coefficient on the interaction suggests
that firms that avoid taxes and have high relative internal resources invest more than
firms that avoid taxes and have less relative internal resources; supporting the notion that
overinvestment is the likely source of the positive relation.
Third, I partition on the volatility of a firm’s tax avoidance. I follow McGuire et
al. (2011a) by measuring the volatility of tax avoidance using the coefficient of variation
over the five year period in which I measure tax avoidance.25
Specifically, CVCETR5
equals the standard deviation of annual cash ETRs from year t-6 to t-1 divided by the
absolute value of the mean of annual cash ETR over the same five-year period.
CVCETR5 is increasing in the volatility of tax avoidance, controlling for the average
level of tax avoidance. I interact CVCETR5 with negCETR5. A negative coefficient on
the interaction suggests that firms that have high volatility in their tax avoidance rely
less on tax avoidance to fund investment, consistent with firms facing the risk of
suffering a shortfall in cash savings from tax avoidance not relying on tax avoidance to
alleviate capital rationing. A positive coefficient on the interaction suggests firms that
avoid taxes, but do so inconsistently, overinvest their cash flow windfalls from tax
avoidance.
25
McGuire et al. (2011a) refer to their measure as the sustainability of tax avoidance. I refer to their
measure as the volatility of tax avoidance to conceptually link my research question with prior literature
on the volatility of cash flows.
26
Controls
I control for other determinants of investments that have been cited in the
literature. Biddle et al. (2009) and McNichols and Stubben (2008) find that financial
reporting quality decreases investment inefficiency. I control for financial reporting
quality using one of two proxies. I follow Biddle et al. (2009) and use the cross sectional
Dechow and Dichev (2002) model supplemented with variables suggested in McNichols
(2002). I measure Dechow as the standard deviation of the previous five years’ residuals
(ei,t) from Model (6). I multiply Dechow by negative one for ease of interpretation, such
that higher values of Dechow indicate better financial reporting quality. I also follow
Ayers et al. (2009) and use the absolute value of discretionary accruals calculated with
the modified Jones (1991) model. AbsDACC is the absolute value of the residual (ui,t)
from Model (7). I estimate both Models (6) and (7) for every industry-year with at least
20 valid observations.26
Accrualsi,t = c0 + c1 CFi,t-1 + c2 CF i,t + c3 CF i,t+1 + c4 dRevi,t + c5 PPEi,t + ei,t (6)
Accrualsi,t = d0 + d1 1/Ai,t-1 + d2(dRevi,t -dReci,t) + d3 PPEi,t + ui,t (7)
where:
Accrualsi,t-1 =The difference between income before extraordinary items and
operating cash flows scaled by lagged total assets [(IB-
OANCF)/AT]
CF = Operating cash flow scaled by lagged total assets (OANCF/AT)
26
Requiring only ten valid observations for an industry-year only increases my sample size by 426 firm-
year observations. Results are robust to the inclusion of these additional observations.
27
dRev = Change in revenue (SALE) from year t-1 to year t scaled by
average total assets (AT)
dRec = Change in receivables (RECT) from year t – 1 to year t scaled
by average total assets (AT)
1/A = The invest of lagged total asset (AT)
PPE = Gross property, plant, and equipment scaled by lagged assets
(PPEGT/AT)
Because deviations from expected investment are the result of capital market
imperfections, I include proxies for corporate governance and the firm’s information
environment based on prior research that links firms operating and investing decisions to
these variables (Core et al. 2006; Ferreira and Matos 2008). IO is the percentage of
shares held by institutional ownership from Thompson Reuters. To control for a firm’s
information environment, I include analyst following (Numest) from I/B/E/S. InvG is the
inverse of a firm’s G-index from Risk Metrics. Due to Risk Metrics issuing the G-index
biannually, I assume that a firm’s G-index in year t is the same in year t+1 (with an
updated score being issued in t+2). Because not all sample firms have a G-index, I create
an indicator variable Ginddum, which is set equal to one to indicate that a firm does not
have a G-index.27
Approximately two-thirds of my sample lacks a G-index. For firms
with missing G-indexes, I follow Biddle et al. (2009) and replace their G-index with
zero. Omitting such a large portion of my sample would limit the generalizability of my
27
I follow Biddle et al. (2009) in setting missing G-index to zero and creating a variable for a missing G-
index. As a robustness test, I follow an alternative data imputation method and set all missing variables
equal to the sample average (-9.21). Results are inferentially similar.
28
findings. I use both InvG and Ginddum in my models so I can control for the effect of
governance without unnecessarily limiting my sample size. I predict firms with greater
IO, InvG, and Numest engage in less overinvestment as monitoring and governance
should alleviate both the moral hazard and adverse selection problems.
I expect firms with greater leverage to invest less than firms with less leverage
because returns to investment may accrue to debtholders and not benefit management or
shareholders (Myers 1977). I calculate leverage as the ratio of long-term debt (DLTT) to
assets (AT). Firms with higher levels of cash are less reliant on external financing and
more likely to overinvest (Blanchard et al. 1994; Jensen 1986). I measure Cash as the
ratio of cash and short term investments (CHE) to assets (AT). LnAT is the natural log of
a firm’s assets (AT) and proxies for size. Larger firms are likely to have fewer capital
constraints and more sophisticated internal controls and thus be less prone to under- and
overinvestment (Biddle et al. 2009). MTB is a firm’s market-to-book ratio, calculated as
the ratio of the market value of equity (PRCC_F*CSHO) scaled by the book value of
equity (CEQ). As prior literature finds that MTB is positively correlated with growth
opportunities, I expect firm with higher values of MTB to have larger investment.
I include StdCFO, StdSales, and StdInvest to capture a firm’s operating
environment. Firms in volatile operating environments are likely to face greater
difficulty accessing outside capital (Minton and Schrand 1999). StdCFO is the standard
deviation of a firm’s scaled operating cash flows (OANCF/AT) calculated over the prior
five years. StdSales is the standard deviation of a firm’s scaled sales (SALE/AT)
calculated over the prior five years. StdInvest is the standard deviation of a firm’s scaled
29
investment (Invest/AT) calculated over the prior five years. I require three out of five
observations for StdCFO, StdSales, and StdInvest to be included in my sample.
I measure a firm’s financial distress using the Zscore from Kaplan and Zingales
(1997).28
I expect firms closer to financial distress to invest significantly less. I include
Tang, the ratio of PPE to total assets (PPEGT/AT) to control for a firm’s collateral.
Firms with more collateral have more access to outside debt financing and therefore
invest more than firms with less collateral (Almeida and Campello 2007). Firms with
greater cash flows invest more than firms with less cash flows (Stein 2003). Therefore, I
control for CFOSale, the ratio of OANCF to SALE. Firms that pay dividends have
greater cash flows and therefore are expected to invest less. I include DIV, an indicator
variable equaling one when DV or DVC is greater than zero and zero otherwise.
Age and Opcycle are meant to capture a firm’s business cycle. Mature firms tend to have
fewer investment opportunities. I expect older firms and firms with longer operating
cycles to invest less than younger firms and firms with shorter operating cycles. I
measure operating cycle as the natural log of the sum of receivables turnover
(RECT/SALE) and inventory turnover (COGS/INVT), multiplied by 360. I measure age
as the number of years since a firm was first reported in CRSP. Finally, I control for
firms with losses (when IB < 0) as firms with losses have less internal resources to invest
(Stein 2003).
28
I calculate Zscore as (3.3*PI+SALE+.25*RE-.5(ACT-LCT))/AT.
30
CHAPTER IV
SAMPLE SELECTION
To develop my various subsamples, I select all Compustat firms between 1992
and 2008 that have positive asset values, sales, and PPE, are incorporated in the United
States, and have non-negative values of capital expenditure (CAPX), research and
development (XRD), cash acquisitions (AQC), and sales of property, plant, and
equipment (SPPE) (129,293 firm-years).29
The sample period begins in fiscal year 1993
because ASC 740 became effective for fiscal years beginning after December 15th
, 1992,
ensuring consistent reporting for income taxes throughout the entire sample period.30
The sample period ends in 2008 to allow for a one year lead value to calculate the
dependent variable. I remove financial firms as their investment behavior is likely to be
constrained by regulations (119,352 observations).
I then limit the sample to firms with at least 20 industry-year observations to
estimate Model (2) (96,263 firm-years remaining). I define industries using the Fama-
French 48 industry definitions. Because the moral hazard tests require analysis on
truncated samples, I require observations to have similarly-signed residuals across the
Tobin’s Q and sales growth versions of Model (2) (90,009 observations remaining).
29
FIN 48 became effective in 2007. My sample therefore contains two years in which FIN 48 was
effective. To allow for the possibility that FIN 48 changed the relation between tax avoidance and
investment, I include an indicator variable that equals one for years following FIN 48 and interact it with
negCETR5. Both the main effect and the interaction are insignificant, suggesting that FIN 48 did not alter
the relation between tax avoidance and investment. 30
As fiscal years and calendars years are not perfectly correlated, an immaterial portion of my sample has
fiscal years ending before December 15th
, 1992 (26 observations) and are thus reporting their taxes under
APB 11 rather than ASC 740. Results are inferentially similar when removing these observations.
31
Observations with conflicting signs are not clearly identifiable as either investing above
or below an expected level and likely contain excessive measurement error.
I eliminate firms that do not have adequate information for the control variables
specified in Model (1). Because earnings quality, cash flow, and sales reporting were
unlikely to have been affected by SFAS 109, I use information prior to 1992 in order to
calculate Dechow, StdCFO, and StdSales. I obtain institutional ownership data from
Thompson Reuters, Numest from the I/B/E/S summary file, and G-indexes from Risk
Metrics. I follow Biddle et al. (2009) and assume that if a firm has missing institutional
ownership or missing analyst following then these variables are equal to zero.31
After
eliminating firms because of these constraints, the sample has 57,153 firm-year
observations. Next, I require firms to have a nonmissing, valid value for negCETR5. I
require firms to have positive values for both the numerator, cash taxes paid (TXPD),
and denominator, pretax income less special items (PI less SPI) because I am not able to
determine if firms receiving refunds or experiencing losses are avoiding taxes. This
produces my final samples: 30,232 observations for models using Dechow and 29,585
observations for models AbsDACC.
31
Missing institutional ownership occurs for approximately 16 percent of my sample. Approximately 18
percent of the sample lacks analyst following.
32
CHAPTER V
RESULTS
Univariate Statistics
Table 1 contains univariate statistics for my sample. Approximately one percent
of observations have negative investment, indicating asset sales in excess of new
investment. Approximately 65 percent of my sample invests below their expected level.
This distribution of unexpected investment is similar to the findings of Biddle et al.
(2009) and Verdi (2006). My measures of tax avoidance are slightly higher than Dyreng
et al. (2008). My sample has an average negCETR5 of -31.1 percent. On average, Tax
Cash Savings is 0.006, suggesting that, on average, firms save 0.6 percent of assets
through tax avoidance.
Correlations
Table 2 examines the correlations between tax avoidance and investment.
Pearson correlation coefficients are above the main diagonal and Spearman correlation
coefficients are below the main diagonal. I find a positive correlation between both
measures of investment (p-value < 0.001), suggesting firm’s investment responds to
investment opportunities as suggested by Hayashi (1982) and Tobin (1969). Finally, I
find significantly positive correlations between negCETR5 and both investment
measures (p-value < 0.001). However, I am unable to fully determine this relation
without the inclusion of additional control variables.
33
Multivariate Results
I estimate Model (2) by industry-year. As expected, sales growth is positively
associated with a firm’s investment (p-value < 0.001). Of the 763 industry-year groups,
sales growth is significantly positive in 262 specifications and significantly negatively in
only 14 industry-years (untabulated).
Table 3 presents the results of estimating Model (1); the dependent variable of
Columns (1) and (2) is Invest while UnExp_Invest is the dependent variable of Columns
(3) and (4). Even-numbered columns control for financial reporting quality using
Dechow while odd-numbered columns use AbsDACC to control for financial reporting
quality. The coefficients on negCETR5 are significantly positive, indicating that, on
average, firms’ investment increases with tax avoidance (p-value < 0.001). I fail to reject
H1 as I find a positive association between tax avoidance and firm-level investment.
The control variables in Table 3 are broadly consistent with expectations and
prior research. Firms with better financial reporting quality (Dechow and AbsDACC)
invest less than firms with worse financial reporting quality (McNichols and Stubben
2008). Similarly, firms with fewer antitakeover provisions invest less than firms with
more antitakeover provisions (InvG) (Biddle et al. 2009; Khurana et al. 2011). Firms
with greater institutional ownership and greater cash, on average, invest significantly
more than firms with less institutional ownership or cash (Biddle et al. 2009; Richardson
2006). This suggests firms with more internal resources and greater access to external
capital are able to invest more. Also, the market-to-book ratio and cash flow volatility
are significantly and positively associated with investment as in Biddle et al. (2009) and
34
Richardson (2006). Firms with higher leverage and that are closer to financial distress
invest significantly less, consistent with Myers (1977). StdSales is significantly negative,
suggesting firms in more volatile consumer markets invest less than firms in steadier
consumer markets (Biddle et al. 2009; Khurana et al. 2011). StdInvest is significantly
positive, as in Biddle et al. (2009). Tang is significantly positive, consistent Almeida
and Campello (2007). Firms with longer operating cycles as well as older firms invest
less than firms with shorter operating cycles and younger firms, consistent with prior
research indicating that the operating environment and life cycle of firms affecting
investment decisions (Biddle et al. 2009; Khurana et al. 2011). CFOSale is significantly
negative, consistent with prior research (Biddle et al. 2009; Khurana et al. 2011). Firms
with losses and firms that pay dividends invest significantly less, consistent with firms
with less internal resources investing less than firms with more internal resources.
Table 4 presents my results of estimating Equation (5). The dependent variable in
all columns is Invest. Columns (1) and (2) include my full sample of firms while
Columns (3) and (4) limit my sample to firms with positive pre-tax income. I continue to
fail to reject H1 as the coefficients on Tax Cash Savings are significant and positive in
all columns (p-value = 0.002, 0.010, 0.000, and 0.000, respectively). On average, firms
invest between 10.3-cents and 17.9-cents of one additional dollar of cash savings from
tax avoidance. Firms invest between 20.0-cents and 22.4-cents of every additional dollar
of pre-tax cash flows. The coefficients on PTCF as significantly greater than Tax Cash
Savings in both columns, with p-values equaling 0.098 and 0.000, respectively. In all but
Column (3), firms invest a significantly greater portion of pretax cash flows than cash
35
savings from tax avoidance (p-value = 0.098, 0.029, and 0.087, respectively). In
Columns (3), the coefficients for Tax Cash Savings and PTCF are statistically
indistinguishable with p-value = 0.398.
Determining the Source of the Positive Relation
The positive relation documented in the previous section is consistent with
adverse selection’s capital rationing and moral hazard’s overinvestment being the source
of the positive relation between tax avoidance and investment. To distinguish between
these sources, I partition the sample on CEO ownership, relative internal resources, and
the volatility of tax avoidance because each of these variables has unique predictions
associated with either capital rationing or overinvestment.
The results for CEO ownership are presented in Table 5.32
The dependent
variable in Columns (1) and (2) is Invest while the last two columns’ dependent variable
is UnExp_Invest. A positive coefficient is consistent with firms that face greater adverse
selection problems, and therefore greater capital rationing, relying on tax avoidance to
fund investments whereas a negative coefficient is consistent with firms with less moral
hazard problems overinvesting less. Although negCETR5 is only significant in Column
(2), the coefficients for tax avoidance are significantly positive when interacted with the
top quintile of CEO ownership (p-value = 0.050, 0.041, 0.053, and 0.038 respectively). I
therefore reject H2 in support of a positive interaction between tax avoidance and high
CEO ownership. Ownership appears to encourage, rather than discourage, managers to
use tax avoidance as a form of internal financing. This is consistent with capital
32
The results presented here are only for CEO ownership. Results also hold when I use total ownership of
the top five managers as well as the average ownership of the top five managers.
36
rationing being the likely source of the positive relation between tax avoidance and
investment and inconsistent with overinvestment.
Table 6 presents the results from my high relative internal resource tests. The
dependent variable of the first two columns is Invest while the dependent variable for the
remaining two columns is UnExp_Invest. A negative coefficient on the interaction of
negCETR5 and High_RIR is consistent with firms that require less funding for
investment opportunities (and therefore less likely to forgo investment because of capital
rationing) not relying on tax avoidance to fund investments whereas a positive
coefficient is consistent with firms with more moral hazard problems overinvesting
more. The coefficient for negCETR5 continues to be significantly positive with p-values
less than 0.001 in all four specifications. I reject H3 as the coefficients on
negCETR5_x_HighRIR are negative and significant (p-values of 0.002, 0.005, 0.001, and
0.005, respectively). Overall, these findings suggest firms with internal resources that are
relatively greater than investment opportunities relying less on tax avoidance to fund
investment and inconsistent with overinvestment.
Finally, I interact tax avoidance with the volatility of tax avoidance in Table 7. A
negative coefficient is consistent with firms with more volatile tax avoidance not being
able to fund investment opportunities due to cash tax savings shortfalls and not relying
on tax avoidance to fund investments whereas a positive coefficient is consistent with
firms with greater volatility in tax avoidance overinvesting the cash windfalls. The
coefficient on negCETR5 is positive and significant in Columns (1) through (4) (p-
values < 0.001). The coefficients on the interactions between negCETR5 and CVCETR5
37
are also significant with p-values equaling than 0.003, 0.007, 0.003, and 0.007,
respectively. I therefore reject H4 in all columns as the coefficients on the interaction of
negCurETR5 and High_RIR are negative and significant. This is consistent with firms
that cannot consistently rely upon cash savings from tax avoidance not relying on it to
fund investments and is inconsistent with overinvestment.33
33
As noted in footnote 13, the volatility of operating cash flows likely exacerbates both capital rationing
and overinvestment and thus is not useful in distinguishing the underlying source of the positive relation
between tax avoidance and investment. However, the volatility of operating cash flows provides a useful
setting to triangulate my results and corroborate my findings. As firms with highly volatile operating cash
flows are the most likely to be experiencing capital rationing or overinvestment, the interaction between
negCETR5 and CVCETR5 should be significantly greater in magnitude compared to firms with less
volatile operating cash flows. In untabulated results, I estimate the models in Table 7 for firms in the
highest quintile of operating cash flow volatility and firms in the lowest quintile of operating cash flow
volatility. The interaction between negCETR5 and CVCETR5 is significantly negative only for firms with
the highest volatility of operating cash flow and insignificant among firms with the smoothest operating
cash flows. This is consistent with firms that experience the greatest capital rationing relying on tax
savings to finance investment and more so if they sustainable tax savings.
38
CHAPTER VI
ROBUSTNESS TESTS
I perform several robustness tests to verify my findings. First, some forms of
investments have tax preferences that might lead to a mechanical association between
tax avoidance and investment. Two tax preferences are salient in my setting: the research
and development tax credit and the accelerated depreciation of personal (i.e. non-real)
assets. To the extent that firms respond to these tax incentives, firms increase their
investments in order to receive the tax benefits and decrease the firm’s cash taxes paid.
As a result, some current period capital expenditures and research and development and
current period tax avoidance are mechanically related. Although the tax avoidance levels
are lagged in the primary test, if investment is serially-correlated, then lagged tax
avoidance measures may be mechanically related to research and development or to
capital expenditures.
I ensure the results are not due to the research and development tax credit by
limiting the sample to firms without research and development activities (where XRD is
missing in Compustat). By definition, these firms’ investments include only capital
expenditures and acquisitions and are therefore ineligible for the research and
development tax credit. Table 8 presents the results of estimating Model (1) over firms
that do not invest in research and development. The dependent variable in Columns (1)
and (2) is Invest while UnExp_Invest is the dependent variable in Columns (3) and (4).
The coefficients on negCETR5 are significantly positively in all four columns,
39
suggesting my primary results are not a result of the research and development credit (p-
value < 0.01).
Next, I consider accelerated depreciation by measuring tax avoidance from year
t-7 to year t-2, providing an additional year’s separation between the measurement of tax
avoidance and when a firm’s investment occurs. By including an additional year’s
separation, I further minimize the possibility that the results are due to a mechanical
association due to accelerated tax depreciation. Table 9 presents the results of estimating
Model (1) with twice lagged tax avoidance. Invest is the dependent variable in Columns
(1) and (2) and UnExp_Invest is the dependent variable in Columns (3) and (4). I
continue to find significant and positive coefficients on negCETR5 in all four columns,
suggesting that accelerated depreciation is not likely to influence my primary findings
(p-value < 0.01).
In addition, I draw on Richardson’s (2006) methodology to ensure my findings
are not limited to my variable and model specifications. I make two changes to the
primary analysis to follow Richardson (2006) and Blaylock (2011). First, Richardson
(2006) measures investment net of depreciation (DPC). Depreciation is a proxy for the
expected amount of investment necessary to maintain assets in place. For investment to
be considered to increase a firm’s capital stock, it must exceed the present year’s
depreciation. Second, Richardson (2006) and Blaylock (2011) include profitability, cash
holdings, age, leverage, size, and lagged investment in their first stage regressions.
Following Richardson (2006), I include these additional explanatory variables in Model
(2) to ensure that my tax avoidance measure is not capturing the effects of previously
40
documented determinants of investment. I estimate this augmented first stage model by
industry-year and use the residuals from this model as a measure of unexpected
investment. I then estimate Model (1), using the residuals from the Richardson (2006)
model as the dependent variable (Invest_Rich) and omitting the variables previously
included in the first stage.
Coefficients for the first stage model are consistent with expectations
(untabulated). Firms with greater cash, greater prior period investment, greater size,
more profitability, and greater investment opportunities invest significantly more, ceteris
paribus. All else equal, older firms and firms with higher leverage invest less than older
firms and lower leverage. The results are consistent with results from Richardson’s
(2006) and Blaylock’s (2011) first-stage regressions.
Table 10 presents the results of using Invest_Rich as the dependent variable. I
continue to find the coefficients on negCETR5 are positive and significant in Columns
(1) and (2), indicating tax avoidance and investment are positively associated. This
positive relation is consistent with capital rationing and overinvestment.
My next robustness test is based on Bergstresser (2006), who notes that a
positive relation between internal resources and investment need not imply a capital
market imperfection. If internal resources are positively correlated with investment
opportunities, then regressing investment on internal resources would yield a positive
coefficient. However, the positive coefficient may only reflect “the joint movement in
investment and cash flow caused by their correlation with investment opportunities” and
not a capital market imperfection (Bergstresser 2006). If firms with more investment
41
opportunities engage in more tax avoidance than firms with poor investment
opportunities, then my primary results may simply reflect this co-movement and not
indicate tax avoidance being used as a form of internal financing. To correct for any
correlation between tax avoidance activities and investment opportunities, I regress
negCETR5 on sales growth and use the residual (Res_C5) as a measure of tax avoidance
that is independent of investment opportunities.34
I find that investment opportunities are
positively and significantly related to tax avoidance, suggesting that tax avoidance
contains information about investment opportunities.
I tabulate the results of this analysis in Table 11. The dependent variable in
Columns (1) and (2) is Invest while the dependent variable in Columns (3) and (4) is
UnExp_Invest. I continue to find positive and significant coefficients on Res_C5 in all
four columns (p-value < 0.01). These findings suggest my results are not because of
investment opportunities being a correlated, omitted variable with tax avoidance.
Finally, I follow Rauh’s (2006) methodology to determine if the results hold in a
classical investment-sensitivity-to-cash-flow methodology. I regress investment on
Tobin’s Q, cash flow, and tax avoidance, along with firm and year fixed effects.35
A
fixed effect design has the benefit of controlling for unobservable firm-level
characteristics that determine investment (Wooldridge 2002).
34
The extent to which foreign operations represent growth opportunities and also influence tax avoidance,
my results might simply be an artifact of firms with foreign operations investing overseas in low tax
jurisdictions. I address this concern by orthogonalizing tax avoidance on sales growth and an indicator
variable for foreign operations and using the residual as a proxy for tax avoidance. Results are inferentially
similar to those presented in Table 12. 35
I rely on fixed effects over first differences for two reasons. First, when exogeneity assumptions are met,
fixed effects are a more efficient estimator (Wooldridge 2002). Second, first-differencing cash taxes paid
does not provide a clear indicator of the change in tax avoidance as one year measures have significant
variation and measurement error (Dyreng et al. 2008).
42
I present the results of this analysis in Table 12. The dependent variable in
Column (1) is Invest and UnExp_Invest is the dependent variable in Column (2).
Consistent with Rauh (2006), I find positive and significant coefficients on Tobin’s Q
and cash flow in both columns (p-value < 0.01). I also find positive and significant
coefficients for negCETR5 in both columns (p-value < 0.01), suggesting that as firm-
level tax avoidance increases above the firm-level average, investment increases.
However, I do not rely heavily on this modification of investment to cash flow
sensitivity methodology given that Kaplan and Zingales (1997) find that investment-to-
cash-flow sensitivities are high for many subsamples of more financially constrained
firms compared to firms that are less financially constrained.
43
CHAPTER VII
CONCLUSION
Tax avoidance, as well as other forms of internal resources, can theoretically be
related to firm-level investment because of two capital market imperfections: adverse
selection and moral hazard (Jensen 1986; Jensen and Meckling 1976; Myers and Majluf
1984; Stein 2003). Using these two capital market imperfections to provide theoretical
links to investment, I suggest two expectations for a relation between tax avoidance and
investment, both of which predict a positive relation. First, tax avoidance can be a form
of internal financing to overcome adverse selection’s capital rationing. Second,
managers can also use the obfuscation associated with tax avoidance to grow the size of
the firm beyond what is optimal by overinvesting, another form of moral hazard (Desai
and Dharmapala 2006).
I provide robust evidence that tax avoidance is positively related to investment,
suggesting adverse selection’s capital rationing or overinvestment are potential sources
of the positive relation. Further, I provide evidence that managerial ownership
encourages managers to rely on tax avoidance as a form of internal financing. This is
consistent with firms with CEOs that have greater incentives to exploit potential, future
shareholders, relying on the cash savings from tax avoidance to alleviate capital
rationing. I also find that firms with high relative internal resources rely on tax
avoidance to fund investments to a significantly lesser extent than firms with low
relative internal resources. This is consistent with capital rationing being the likely
44
source of the positive relation and suggests that firms that are less likely to forgo
profitable investments because of capital rationing requiring less of the cash savings of
tax avoidance to fund investment. I also find evidence that firms with volatile tax
avoidance strategies rely on tax avoidance to a significantly lesser extent to fund
investment than firms with less volatile tax avoidance strategies, consistent with capital
rationing being the source of the positive relation between tax avoidance and investment.
Taken together, my findings support the notion that tax avoidance increases
investment by increasing internal resources and therefore alleviating adverse selection’s
capital rationing. The results provide one possible mechanism, overcoming financing
constraints, whereby tax avoidance increases firm value (Koester 2011; Wang 2010;
Wilson 2009). Overall, my findings suggest that shareholders should view tax
avoidance, at least on average, as beneficial to the firm as it increases positive-NPV
investments. Moreover, my findings are important to policymakers who need to
understand the benefits, as well as the costs, of legislations seeking to curtail tax
avoidance strategies and managers seeking to overcome capital rationing.
45
REFERENCES
Allen, J. W., and G. M. Phillips. 2000. Corporate Equity Ownership, Strategic Alliances,
and Product Market Relationships. Journal of Finance 55 (6):2791-2815.
Allingham, M. G., and A. Sandmo. 1972. Income Tax Evasion: A Theoretical Analysis.
Journal of Public Economics 1 (3/4):323.
Almeida, H., and M. Campello. 2007. Financial Constraints, Asset Tangibility, and
Corporate Investment. Review of Financial Studies 20 (5):1429-1460.
Ayers, B. C., J. Jiang, and S. K. Laplante. 2009. Taxable Income as a Performance
Measure: The Effects of Tax Planning and Earnings Quality. Contemporary
Accounting Research 26 (1):15-54.
Babenko, I., M. Lemmon, and Y. Tserlukevich. 2011. Employee Stock Options and
Investment. Journal of Finance 66 (3):981-1009.
Badertscher, B., S. P. Katz, and S. O. Rego. 2010. The Impact of Public and Private
Equity Ownership, Financial Reporting Pressure, and Debt Financing on
Corporate Tax Avoidance. University of Notre Dame Working Paper.
Balakrishnan, K., J. L. Blouin, and W. R. Guay. 2011. Does Tax Aggressiveness Reduce
Financial Reporting Transparency? University of Pennsylvania Working Paper.
Bergstresser, D. 2006. Discussion of “Overinvestment of Free Cash Flow”. Review of
Accounting Studies 11 (2/3):191-202.
Biddle, G. C., and G. Hilary. 2006. Accounting Quality and Firm-Level Capital
Investment. Accounting Review 81 (5):963-982.
Biddle, G. C., G. Hilary, and R. S. Verdi. 2009. How Does Financial Reporting Quality
Relate to Investment Efficiency? Journal of Accounting & Economics 48
(2/3):112-131.
Blanchard, O. J., F. Lopez-de-Silanes, and A. Shleifer. 1994. What do firms do with cash
windfalls? Journal of Financial Economics 36 (3):337-360.
Blaylcok, B. 2011. Is Tax Avoidance Associated with Managerial Rent Extraction from
Shareholders Among Poorly Governed US Firms? Oklahoma State University
Working Paper.
46
Chen, S., X. Chen, Q. Cheng, and T. Shevlin. 2010. Are Family Firms More Tax
Aggressive than Non-family Firms? Journal of Financial Economics 95 (1):41-
61.
Core, J. E., W. R. Guay, and T. O. Rusticus. 2006. Does Weak Governance Cause Weak
Stock Returns? An Examination of Firm Operating Performance and Investors'
Expectations. Journal of Finance 61 (2):655-687.
Dechow, P. M., and I. D. Dichev. 2002. The Quality of Accruals and Earnings: The Role
of Accrual Estimation Errors. Accounting Review 77 (4):35.
Desai, M. A., and D. Dharmapala. 2006. Corporate Tax Avoidance and High-Powered
Incentives. Journal of Financial Economics 79 (1):145-179.
———. 2009. Corporate Tax Avoidance and Firm Value. Review of Economics &
Statistics 91 (3):537-546.
Dhaliwal, D. S., S. X. Huang, W. J. Moser, and R. Pereira. 2011. Corporate Tax
Avoidance and the Level and Valuation of Firm Cash Holdings. University of
Missouri-Columbia Working Paper.
Dyreng, S. D., M. Hanlon, and E. L. Maydew. 2008. Long-Run Corporate Tax
Avoidance. Accounting Review 83 (1):61-82.
Fazzari, S. M., R. G. Hubbard, and B. C. Petersen. 1988. Financing Constraints and
Corporate Investment. Brookings Papers on Economic Activity (1):141-206.
Ferreira, M. A., and P. Matos. 2008. The Colors of Investors’ Money: The Role of
Institutional Investors Around the World. Journal of Financial Economics 88
(3):499-533.
Frank, M. M., L. J. Lynch, and S. O. Rego. 2009. Tax Reporting Aggressiveness and Its
Relation to Aggressive Financial Reporting. Accounting Review 84 (2):467-496.
Frischmann, P. J., T. Shevlin, and R. Wilson. 2008. Economic Consequences of
Increasing the Conformity in Accounting for Uncertain Tax Benefits. Journal of
Accounting & Economics 46 (2/3):261-278.
Gaertner, F. B. 2010. CEO After-Tax Compensation Incentives and Corporate Tax
Avoidance. Nanyang Technological University Working Paper.
Gow, I. D., G. Ormazabal, and D. J. Taylor. 2010. Correcting for Cross-Sectional and
Time-Series Dependence in Accounting Research. Accounting Review 85
(2):483-512.
47
Graham, J. R., and A. L. Tucker. 2006. Tax Shelters and Corporate Debt Policy. Journal
of Financial Economics 81 (3):563-594.
Grubert, H. 2003. Intangible Income, Intercompany Transactions, Income Shifting, and
the Choice of Location. National Tax Journal 56 (1):221-242.
Hadlock, C. J. 1998. Ownership, liquidity, and investment. RAND Journal of Economics
29 (3):487-508.
Hadlock, C. J., M. Ryngaert, and S. Thomas. 2001. Corporate Structure and Equity
Offerings: Are There Benefits to Diversification? Journal of Business 74 (4):613-
635.
Hanlon, M. 2003. What Can We Infer about a Firm's Taxable Income from Its Financial
Statements? National Tax Journal 56 (4):831-863.
Hanlon, M., and S. Heitzman. 2010. A Review of Tax Research. Journal of Accounting
& Economics 50 (2/3):127-178.
Hanlon, M., and J. Slemrod. 2009. What Does Tax Aggressiveness Signal? Evidence
from Stock Price Reactions to News about Tax Shelter Involvement. Journal of
Public Economics 93 (1/2):126-141.
Harris, M., and A. Raviv. 1990. Capital Structure and the Informational Role of Debt.
Journal of Finance 45 (2):321-349.
Hart, O., and J. Moore. 1995. Debt and Seniority: An Analysis of the Role of Hard
Claims in Constraining Management. American Economic Review 85 (3):567-
585.
Hayashi, F. 1982. Tobin's Marginal q and Average q: A Neoclassical Interpretation.
Econometrica 50 (1):213-224.
Higgins, D., T. C. Omer, and J. D. Phillips. 2011. Does a Firm s Business Strategy
Influence its Level of Tax Avoidance? University of Connecticut Working Paper.
Hope, O.-K., M. Ma, and W. Thomas. 2012. Tax Avoidance and Geographic Earnings
Disclosure. University of Oklahoma Working Paper.
Hoshi, T., A. Kashyap, and D. Scharfstein. 1991. Corporate Structure, Liquidity, and
Investment: Evidence from Japanese Industrial Groups. Quarterly Journal of
Economics 106 (1):33-60.
48
Hubbard, R. G. 1998. Capital-Market Imperfections and Investment. Journal of
Economic Literature 36 (1):193-225.
Jennings, R., C. D. Weaver, and W. J. Mayew. Forthcoming. The Extent of Implicit
Taxes at the Corporate Level and the Effect of TRA86. Contemporary
Accounting Research.
Jensen, M. C. 1986. Agency Costs of Free Cash Flow, Corporate Finance, and
Takeovers. American Economic Review 76 (2):323.
Jensen, M. C., and W. H. Meckling. 1976. Theory of the Firm: Managerial Behavior,
Agency Costs and Ownership Structure. Journal of Financial Economics 3
(4):305-360.
Jones, J. J. 1991. Earnings Management During Import Relief Investigations. Journal of
Accounting Research 29 (2):193-228.
Kaplan, S. N., and L. Zingales. 1997. Do Investment-Cash Flow Sensitivies Provide
Useful Measures of Financing Constraints? Quarterly Journal of Economics 112
(1):169-215.
Khurana, I., W. J. Moser, and K. Raman. 2011. Does Non-Conforming Tax Avoidance
Affect Investment Decisions. University of Missouri-Columbia Working Paper.
Kim, J.-B., V, Y. Li, and L. Zhang. 2010. Corporate Tax Avoidance and Stock Price
Crash Risk: Firm-Level Analysis. Journal of Financial Economics 100 (3):639-
662.
Koester, A. 2011. Investor Valuation of Tax Avoidance Through Uncertain Tax
Positions. Georgetown University Working Paper.
Kothari, S. P., A. J. Leone, and C. E. Wasley. 2005. Performance matched discretionary
accrual measures. Journal of Accounting & Economics 39 (1):163-197.
Lamont, O. 1997. Cash Flow and Investment: Evidence from Internal Capital Markets.
Journal of Finance 52 (1):83-109.
McGuire, S. T., S. S. Neuman, and T. C. Omer. 2011a. Through the Looking Glass: Are
Sustainable Tax Strategies Reflect in Earnings Persistence? Texas A&M
University Working Paper.
McGuire, S. T., D. D. Wang, and R. J. Wilson. 2011b. Dual Class Ownership and Tax
Avoidance. Texas A&M University Working Paper.
49
McNichols, M. F. 2002. Discussion of The Quality of Accruals and Earnings: The Role
of Accrual Estimation Errors. Accounting Review 77 (4):61.
McNichols, M. F., and S. R. Stubben. 2008. Does Earnings Management Affect Firms'
Investment Decisions? Accounting Review 83 (6):1571-1603.
Mills, L. F. 1998. Book-Tax Differences and Internal Revenue Service Adjustments.
Journal of Accounting Research 36 (2):343-356.
Minnick, K., and T. Noga. 2010. Do Corporate Governance Characteristics Influence
Tax Management? Journal of Corporate Finance 16 (5):703-718.
Minton, B. A., and C. Schrand. 1999. The Impact of Cash Flow Volatility on
Discretionary Investment and the Costs of Debt and Equity Financing. Journal of
Financial Economics 54 (3):423-460.
Minton, B. A., C. M. Schrand, and B. R. Walther. 2002. The Role of Volatility in
Forecasting. Review of Accounting Studies 7 (2/3):195-215.
Modigliani, F., and M. H. Miller. 1958. The Cost of Capital, Corporation Finance, and
the Theory of Investment. American Economic Review 48 (3):261.
Myers, S. C. 1977. Determinants of Corporate Borrowing. Journal of Financial
Economics 5 (2):147-175.
Myers, S. C., and N. S. Majluf. 1984. Corporate Financing and Investment Decisions
when Firms have Information that Investors Do Not Have. Journal of Financial
Economics 13 (2):187-221.
Overesch, M. 2009. The Effects of Multinationals' Profit Shifting Activities on Real
Investments. National Tax Journal 62 (1):5-23.
Ozbas, O., and D. S. Scharfstein. 2010. Evidence on the Dark Side of Internal Capital
Markets. Review of Financial Studies 23 (2):581-599.
Rauh, J. D. 2006. Investment and Financing Constraints: Evidence from the Funding of
Corporate Pension Plans. Journal of Finance 61 (1):33-71.
Rego, S. O., and R. J. Wilson. 2012. Executive Compensation, Equity Risk Incentives,
and Corporate Tax Aggressiveness. Journal of Accounting Research 50 (3):775-
810.
Richardson, S. 2006. Over-investment of Free Cash Flow. Review of Accounting Studies
11 (2/3):159-189.
50
Scholes, M., M. Wolfson, M. Erickson, E. L. Maydew, and T. Shevlin. 2008. Taxes and
Business Strategy: A Planning Approach. Upper Saddle River, NJ: Prentice Hall.
Shackelford, D. A., and T. Shevlin. 2001. Empirical Tax Research in Accounting.
Journal of Accounting & Economics 31 (1-3):321-387.
Stein, J. C. 2003. Agency, Information, and Corporate Investment. edited by M. H.
George Constantinides, Rene Stulz: Elsevier.
Stiglitz, J. E., and A. Weiss. 1981. Credit Rationing in Markets with Imperfect
Information. American Economic Review 71 (3):393.
Stulz, R. M. 1990. Managerial Discretion and Optimal Financing Policies. Journal of
Financial Economics 26 (1):3-27.
Tobin, J. 1969. A General Equilibrium Approach To Monetary Theory. Journal of
Money, Credit & Banking (Ohio State University Press) 1 (1):15-29.
Verdi, R. S. 2006. Financial Reporting Quality and Investment Efficiency. MIT Working
Paper.
Vogt, S. C. 1994. The Cash Flow/Investment Relationship: Evidence from U.S.
Manufacturing Firms. FM: The Journal of the Financial Management
Association 23 (2):3-20.
Wang, X. 2010. Tax Avoidance, Corporate Transparency, and Firm Value. University of
Texas at Austin Working Paper.
Wilson, R. J. 2009. An Examination of Corporate Tax Shelter Participants. Accounting
Review 84 (3):969-999.
Wooldridge, J. 2002. Econometric Analysis of Cross Section and Panel Data.
Cambridge, Massachusetts: The MIT Press.
51
APPENDIX A
VARIABLE DEFINITIONS
Invest The sum of capital expenditure (CAPX), research
and development (XRD), and acquisitions (AQC)
less the sale of property, plant, and equipment
(SPPE) and scaled by lagged total assets (AT) in
accordance with Biddle et al. (2009).
UnExp_Invest The residual from regressing Invest on Sales
Growth
negCETR5 The sum of cash taxes paid (TXPD) from t-5 to t-
1 scaled by pretax income net of special items (PI
– SPI) over the same time period and multiplied
by negative one. I winsorize at 0 and 1 in
accordance with Dyreng et al. (2008).
Tax Cash Savings The difference between 35% and CETR5,
multiplied by pretax income (PI) and scaled by
lagged total assets (AT)
PTCF Operating cash flows (OANCF) plus cash taxes
paid (TXPD), scaled by lagged total assets (AT)
HighCEO An indicator variable equaling 1 if the firm is in
the top quintile of CEO ownership and 0
otherwise
High_RIR An indicator variable that equals 1 if a firm’s
Tobin’s Q is in the bottom two quintiles for its
industry and has positive free cash flow and a 0
otherwise. I follow Richardson (2006) by
defining free cash flow as the sum of operating
cash flows (OANCF), research and development
(XRD) minus depreciation (DPC) and the
predicted value from of Invest Model (2).
CVCETR5 The standard deviation of annual cash ETRs
(TXPD/(PI-SPI)) from years t-6 to t-1 divided by
the absolute value of the mean of annual cash
52
ETR over the same five-year period.
Dechow The standard deviation of the prior five years
residual from a regression of accruals on current,
lagged, lead cash flows, the change in revenue,
and property plant and equipment as described in
Dechow and Dichev (2002) and McNichols
(2002). I estimate the model for every industry-
year with twenty valid observations.
IO The percentage of shares held by institutional
ownership from Thompson Reuters
Numest Analyst following from I/B/E/S
InvG The inverse of a firm’s G-index from Risk
Metrics. Since Risk Metrics issues the G-index
biannually, I assume that a firm’s G-index score
in year t is the same in year t+1 (with an updated
score being issued in t+2.
Ginddum An indicator variable equaling 1 to for a firm
does not have a G-index and 0 otherwise.
Lev The ratio of long-term debt (DLTT) to assets (AT)
Cash The ratio of cash and short term investments
(CHE) to assets (AT)
LnAT The natural log of a firm’s assets (AT)
MTB The ratio of the market value of equity
(PRCC_F*CSHO) to the book value of equity
(CEQ)
StdCFO The standard deviation of a firm’s scaled
operating cash flows (OANCF/AT) calculated
over the prior five years
StdSales The standard deviation of a firm’s scaled sales
(SALE/AT) calculated over the prior five years
StdInvest The standard deviation of a firm’s investment
calculated over the prior five years
53
Zscore As calculated in Biddle et al. (2009),
(3.3*PI+SALE+.25*RE-.5*(ACT-LCT))/AT
Tang The ratio of PPE to total assets (PPEGT/AT)
CFOSale The ratio of operating cash flows (OANCF) to
sales (SALE)
DIV An indicator variable equaling 1 if a firm pays
dividends (DV > 0 or DVC > 0) and 0 otherwise.
Age The number of years since a firm was first
reported in CRSP
Opcycle The natural log of the sum of receivables turnover
(RECT/SALE) and inventory turnover
(COGS/INVT) multiplied by 360
Loss An indicator variable equaling 1 if IB is less than
0 and 0 otherwise
AbsDACC The absolute value of discretionary accruals as
described in Kothari et al. (2005). I estimate the
model for every industry-year with twenty valid
observations.
54
APPENDIX B
TABLES
TABLE 1
Univariate Statistics
Variablesa N Mean Std Dev Q1 Median Q3
Invest 33,303 10.936 9.741 4.276 8.237 14.506
UnExp_Invest 33,303 -1.648 9.515 -7.001 -2.959 2.560
negCETR5 30,232 -0.311 0.195 -0.380 -0.298 -0.196
Tax Cash Savings 27,481 0.006 0.002 -0.002 0.003 0.013
Dechow 33,303 -0.053 0.347 -0.059 -0.036 -0.022
AbsDACC 32,427 0.079 0.126 0.020 0.046 0.091
IO 33,303 0.422 0.328 0.083 0.419 0.705
Numest 33,303 7.233 7.972 1.000 5.000 11.000
InvG 33,303 -3.679 4.827 -8.000 0.000 0.000
Ginddum 33,303 0.601 0.490 0.000 1.000 1.000
Lev 33,303 0.168 0.170 0.012 0.134 0.268
Cash 33,303 0.117 0.156 0.015 0.052 0.161
LnAT 33,303 5.889 1.947 4.480 5.832 7.235
MTB 33,303 2.534 2.877 1.221 1.891 3.033
StdCFO 33,303 0.060 0.047 0.028 0.047 0.078
StdSales 33,303 0.184 0.171 0.073 0.132 0.234
StdInvest 33,303 6.703 6.595 2.291 4.481 8.717
Zscore 33,303 1.431 0.897 0.836 1.317 1.868
Tang 33,303 0.310 0.230 0.125 0.247 0.449
CFOSale 33,303 0.092 0.139 0.033 0.077 0.135
DIV 33,303 0.473 0.499 0.000 0.000 1.000
Age 33,303 19.701 16.138 8.000 14.000 27.000
Opcycle 33,303 4.613 0.722 4.225 4.694 5.089
Loss 33,303 0.150 0.357 0.000 0.000 0.000 aSee Appendix for variable definitions.
55
TABLE 2
Correlations
Variables
a Invest UnExp_Invest negCETR5
Invest 1 0.818 *** 0.109 ***
UnExp_Invest 0.698 *** 1 0.061 ***
negCETR5 0.141 *** 0.048 *** 1 aSee Appendix for variable definitions.
56
TABLE 3
Regression of Investment on Tax Avoidance
Ii,t = b0 + b1 negCETR5i,t-1 + Controls’k,t-1 + bj + ei,t
Invest UnExp_Invest
(1) (2) (3) (4)
negCETR5 1.501*** 1.406*** 1.336*** 1.246***
(0.000) (0.000) (0.000) (0.000)
Dechow -0.340 -0.525
(0.726) (0.525)
AbsDACC 5.684*** 5.228***
(0.000) (0.000)
IO 0.801** 0.839** 0.883*** 0.895***
(0.016) (0.011) (0.002) (0.002)
Numest 0.030 0.026 0.031 0.027
(0.146) (0.205) (0.112) (0.165)
InvG -0.127*** -0.131*** -0.154*** -0.158***
(0.003) (0.002) (0.000) (0.000)
Ginddum 1.652*** 1.675*** 1.342*** 1.346***
(0.000) (0.000) (0.001) (0.002)
Lev -13.537*** -13.705*** -12.788*** -12.945***
(0.000) (0.000) (0.000) (0.000)
Cash 2.122** 2.383** 1.887** 2.122**
(0.032) (0.016) (0.041) (0.022)
LnAT 0.318** 0.367*** 0.189 0.231*
(0.012) (0.005) (0.167) (0.096)
MTB 0.318*** 0.309*** 0.258*** 0.250***
(0.000) (0.000) (0.000) (0.000)
StdCFO 8.818*** 7.413*** 8.215*** 6.940***
(0.000) (0.000) (0.000) (0.000)
StdSales -1.921** -2.155*** -2.174*** -2.431***
(0.015) (0.005) (0.004) (0.001)
StdInvest 0.162*** 0.165*** 0.142*** 0.144***
(0.000) (0.000) (0.000) (0.000)
Zscore -1.549*** -1.567*** -1.665*** -1.690***
(0.000) (0.000) (0.000) (0.000)
Tang 12.539*** 12.955*** 12.042*** 12.378***
(0.000) (0.000) (0.000) (0.000)
CFOSale -1.365** -1.344** -1.092** -1.094**
(0.033) (0.031) (0.036) (0.029)
DIV -1.144*** -1.157*** -1.117*** -1.131***
(0.000) (0.000) (0.000) (0.000)
Age -0.047*** -0.049*** -0.046*** -0.047***
(0.000) (0.000) (0.000) (0.000)
OpCycle -1.421*** -1.450*** -1.582*** -1.622***
(0.000) (0.000) (0.000) (0.000)
Loss -2.174*** -2.237*** -1.821*** -1.873***
(0.000) (0.000) (0.000) (0.000)
Constant 14.840*** 10.762*** 4.273*** 5.117**
(0.000) (0.000) (0.009) (0.013)
Observations 30,232 29,585 30,232 29,585
Adjusted R-squared 0.239 0.246 0.221 0.228 p-values are in parentheses (* significant at 10%; ** significant at 5%; *** significant at 1%) I cluster standard errors by firm and year (Gow et al. 2010). Industry fixed-effect coefficients are omitted for brevity. See Appendix
for variable definitions.
57
TABLE 4
Regression of Investment on Tax Cash Savings
t + C s n s t + C t + Controls t
+ + t+ t
Full Sample No Loss
(1) (2) (3) (4)
Tax Cash Savings 0.130*** 0.103** 0.179*** 0.131**
(0.002) (0.014) (0.000) (0.010)
PTCF 0.203*** 0.200*** 0.224*** 0.222***
(0.000) (0.000) (0.000) (0.000)
Dechow -0.023**
(0.015)
absDACC 0.158*** 0.153***
(0.000) (0.000)
Constant 0.231*** 0.210*** 0.327*** 0.309***
(0.000) (0.000) (0.000) (0.000)
Firm Fixed Effects YES YES YES YES
Year Fixed Effects YES YES YES YES
Controls YES YES YES YES
Observations 27,297 26,714 23,499 22,984
Adjusted R-squared 0.479 0.487 0.499 0.506 p-values in parentheses (* significant at 10%; ** significant at 5%; *** significant at 1%)
I cluster standard errors by firm and year (Gow et al. 2010). Firm and year fixed-effect coefficients are omitted for brevity. See Appendix for variable
definitions
58
TABLE 5
Regression of Investment on Tax Avoidance and CEO Ownership
t + n C + C + n C C + Controls + + t
Invest UnExp_Invest
(1) (2) (3) (4)
negCETR5 1.060 1.089* 1.057 1.078
(0.112) (0.092) (0.122) (0.104)
HighCEO 0.246 0.299 0.264 0.345
(0.651) (0.580) (0.615) (0.506)
HighCEO_x_negCETR5 2.871** 2.875** 2.821* 2.875**
(0.050) (0.041) (0.053) (0.038)
Dechow 0.248 0.169
(0.829) (0.887)
AbsDACC 5.885*** 5.638***
(0.000) (0.000)
Constant 21.515*** 24.788*** 10.573*** 9.553***
(0.000) (0.000) (0.000) (0.000)
Controls YES YES YES YES
Observations 14,989 14,639 14,989 14,639
Adjusted R-squared 0.250 0.256 0.243 0.250 p-values are in parentheses (* significant at 10%; ** significant at 5%; *** significant at 1%)
I cluster standard errors by firm and year (Gow et al. 2010). Industry fixed-effect coefficients are omitted for brevity. See Appendix for variable
definitions.
59
TABLE 6
Regression of Investment on Tax Avoidance and Relative Internal Resources
t + n C t
+ + n C t + Controls t
+ + t
Invest UnExp_Invest
(1) (2) (3) (4)
negCETR5 1.748*** 1.645*** 1.570*** 1.471***
(0.000) (0.000) (0.000) (0.000)
High_RIR -1.392*** -1.415*** -0.893*** -0.912***
(0.000) (0.000) (0.001) (0.001)
negCETR5_x_HighRIR -1.976*** -1.849*** -1.959*** -1.829***
(0.002) (0.005) (0.001) (0.005)
Dechow -0.381 -0.570
(0.691) (0.485)
AbsDACC 5.548*** 5.073***
(0.000) (0.000)
Constant 14.729*** 10.551*** 4.202*** 4.984**
(0.000) (0.000) (0.009) (0.016)
Controls YES YES YES YES
Observations 30,017 29,369 30,017 29,369
Adjusted R-squared 0.239 0.246 0.222 0.228 p-values are in parentheses (* significant at 10%; ** significant at 5%; *** significant at 1%)
I cluster standard errors by firm and year (Gow et al. 2010).Industry fixed-effect coefficients are omitted for brevity. See Appendix for variable
definitions.
60
TABLE 7
Regression of Investment on Tax Avoidance and Volatility
t + n C t
+ C C + n C t C C + Controls t
+ + t
Invest UnExp_Invest
(1) (2) (3) (4)
negCETR5 3.567*** 3.253*** 3.387*** 3.101***
(0.000) (0.000) (0.000) (0.000)
CVCETR5 -0.489* -0.447 -0.454* -0.425
(0.087) (0.135) (0.083) (0.121)
negCETR5_x_ CVCETR5 -1.839*** -1.636*** -1.831*** -1.646***
(0.003) (0.007) (0.003) (0.007)
Dechow -0.289 -0.473
(0.765) (0.567)
AbsDACC 5.678*** 5.227***
(0.000) (0.000)
Constant 15.737*** 19.256*** 5.094*** 3.948**
(0.000) (0.000) (0.002) (0.016)
Controls YES YES YES YES
Observations 30,232 29,585 30,232 29,585
Adjusted R-squared 0.237 0.244 0.221 0.228 p-values are in parentheses (* significant at 10%; ** significant at 5%; *** significant at 1%)
I cluster standard errors by firm and year (Gow et al. 2010). Industry fixed-effect coefficients are omitted for brevity. See Appendix for variable
definitions.
61
TABLE 8
Regression of Investment on Tax Avoidance for Firms lacking R&D
Ii,t = b0 + b1 negCETR5i,t-1 + Controls’k,t-1 + bj + ei,t
Invest UnExp_Invest
(1) (2) (3) (4)
negCETR5 1.133*** 1.125*** 1.105*** 1.104***
(0.001) (0.002) (0.001) (0.001)
Dechow -0.397 -0.639
(0.680) (0.452)
AbsDACC 2.149*** 1.821***
(0.006) (0.009)
Constant 14.810*** 13.793*** 5.536** -1.515
(0.000) (0.000) (0.038) (0.529)
Controls YES YES YES YES
Observations 17,585 16,924 17,585 16,924
Adjusted R-squared 0.295 0.302 0.181 0.185 p-values are in parentheses (* significant at 10%; ** significant at 5%; *** significant at 1%)
I cluster standard errors by firm and year (Gow et al. 2010). Industry fixed-effect coefficients are omitted for brevity. See Appendix for variable
definitions.
62
TABLE 9
Regression of Investment on Twice Lagged Tax Avoidance
Ii,t = b0 + b1 negCETR5i,t-2 + Controls’k,t-1 + bj + ei,t
Invest UnExp_Invest
(1) (2) (3) (4)
negCETR5t-2 0.854*** 0.841*** 0.955*** 0.954***
(0.000) (0.000) (0.000) (0.000)
Dechow -0.145 -0.336
(0.891) (0.726)
AbsDACC 2.810*** 2.506***
(0.000) (0.000)
Constant 12.758*** 17.245*** 3.991*** 2.223
(0.000) (0.000) (0.007) (0.247)
Controls YES YES YES YES
Observations 27,570 26,987 27,570 26,987
Adjusted R-squared 0.232 0.236 0.157 0.159 p-values are in parentheses (* significant at 10%; ** significant at 5%; *** significant at 1%)
I cluster standard errors by firm and year (Gow et al. 2010). Industry fixed-effect coefficients are omitted for brevity. See Appendix for variable
definitions
63
TABLE 10
Regression of Unexpected Investment per Richardson (2006) on Tax Avoidance
Invest_Richi,t = b0 + b1 negCETR5i,t-1 + Controls’k,t-1 + bj + ei,t
Invest_Rich
(1) (2)
negCETR5 0.858*** 0.701**
(0.002) (0.012)
Dechow 4.544**
(0.012)
AbsDACC 3.449***
(0.000)
IO -0.197 -0.149
(0.452) (0.573)
Numest -0.031** -0.029**
(0.013) (0.021)
InvG -0.157*** -0.159***
(0.000) (0.000)
Ginddum 1.842*** 1.785***
(0.000) (0.000)
Zscore -1.667*** -1.695***
(0.000) (0.000)
Tang -6.299*** -6.366***
(0.000) (0.000)
CFOSale -1.531*** -1.441***
(0.007) (0.008)
DIV 0.002 0.034
(0.990) (0.819)
OpCycle -2.213*** -2.280***
(0.000) (0.000)
Loss -2.901*** -3.017***
(0.000) (0.000)
Constant 15.351*** 14.120***
(0.000) (0.000)
Observations 30,141 29,544
Adjusted R-squared 0.065 0.068 p-values are in parentheses (* significant at 10%; ** significant at 5%; *** significant at 1%)
I cluster standard errors by firm and year (Gow et al. 2010). Industry fixed-effect coefficients are omitted for brevity. See Appendix for variable definitions.
64
TABLE 11
Regression of Investment on Tax Avoidance Orthogonalized on Growth Opportunities
Ii,t = b0 + b1 Res_C5i,t-1 + Controls’k,t-1 + bj + ei,t
Invest UnExp_Invest
(1) (2) (3) (4)
Res_C5 1.760*** 1.717*** 1.930*** 1.881***
(0.000) (0.000) (0.000) (0.000)
Dechow -0.379 -0.700
(0.722) (0.450)
AbsDACC 5.705*** 5.246***
(0.000) (0.000)
IO 0.827** 0.866*** 0.914*** 0.926***
(0.013) (0.009) (0.001) (0.001)
Numest 0.029 0.025 0.027 0.024
(0.170) (0.220) (0.158) (0.209)
InvG -0.122*** -0.126*** -0.152*** -0.155***
(0.005) (0.004) (0.000) (0.000)
Ginddum 1.633*** 1.657*** 1.314*** 1.317***
(0.001) (0.000) (0.002) (0.002)
Lev -13.452*** -13.623*** -12.772*** -12.932***
(0.000) (0.000) (0.000) (0.000)
Cash 2.210** 2.467** 1.936** 2.166**
(0.027) (0.014) (0.037) (0.021)
LnAT 0.323** 0.372*** 0.195 0.237*
(0.012) (0.005) (0.154) (0.089)
MTB 0.320*** 0.308*** 0.260*** 0.250***
(0.000) (0.000) (0.000) (0.000)
StdCFO 8.671*** 7.261*** 8.027*** 6.794***
(0.000) (0.000) (0.000) (0.000)
StdSales -1.894** -2.120*** -2.144*** -2.389***
(0.020) (0.008) (0.005) (0.002)
StdInvest 0.165*** 0.168*** 0.145*** 0.147***
(0.000) (0.000) (0.000) (0.000)
Zscore -1.533*** -1.545*** -1.653*** -1.671***
(0.000) (0.000) (0.000) (0.000)
Tang 12.461*** 12.871*** 11.907*** 12.234***
(0.000) (0.000) (0.000) (0.000)
CFOSale -1.355** -1.329** -1.136** -1.132**
(0.031) (0.030) (0.026) (0.021)
DIV -1.144*** -1.161*** -1.118*** -1.138***
(0.000) (0.000) (0.000) (0.000)
Age -0.047*** -0.049*** -0.045*** -0.047***
(0.000) (0.000) (0.000) (0.000)
OpCycle -1.423*** -1.455*** -1.595*** -1.638***
(0.000) (0.000) (0.000) (0.000)
Loss -2.152*** -2.202*** -1.807*** -1.846***
(0.000) (0.000) (0.000) (0.000)
Constant 13.334*** 13.119*** 4.764*** 4.886***
(0.000) (0.000) (0.002) (0.001)
Observations 29,612 28,942 29,612 28,942
Adjusted R-squared 0.240 0.246 0.222 0.229 p-values are in parentheses (* significant at 10%; ** significant at 5%; *** significant at 1%) I cluster standard errors by firm and year (Gow et al. 2010). Industry fixed-effect coefficients are omitted for brevity. See Appendix
for variable definitions.
65
TABLE 12
Regression of Investment on Tax Avoidance using the Rauh (2006) Methodology
Ii,t = b0 + b1 negCETR5i,t-1 + Controls’k,t-1 + bi + bt + ei,t
Invest UnExp_Inv
(1) (2)
negCETR5 0.018*** 1.234***
(0.002) (0.001)
TQ 0.023*** 1.060***
(0.000) (0.000)
PTCF 0.189*** 3.299***
(0.000) (0.000)
Constant 0.067*** -1.550***
(0.000) (0.004)
Firm Fixed Effects YES YES
Year Fixed Effects YES YES
Observations 27,452 27,452
Adjusted R-squared 0.313 0.382 p-values are in parentheses (* significant at 10%; ** significant at 5%; *** significant at 1%)
I cluster standard errors by firm and year (Gow et al. 2010). Industry fixed-effect coefficients are omitted for
brevity. See Appendix for variable definitions.